An intercomparison study on models of sensible heat flux over partial canopy surfaces with remotely sensed surface temperature

Using remotely sensed surface temperature to estimate the sensible heat flux over partial canopy covered surfaces, one faces the problem of how the bare soil and plant foliage temperatures contributing to the radiometric surface temperature are related to the turbulent transport of sensible heat across the surface-atmosphere interface. To solve this problem, several sensible heat models, using radiometric surface temperature, have appeared in the literature. In this study, using the observational data from three interdisciplinary field experiments [the First International Satellite Land Surface Climatology Project (ISLSCP) Field Experiment (FIFE), Monsoon ʹ90 and Washita ʹ92, the performance of four models using average values of their parameters in predicting land surface sensible heat flux was evaluated. By analyzing the sensitivity of the models to common parameters and input variables, reasons for differences in the performances of the models and the potential to improve the agreement with observations have been ascertained. From the comparisons of modeled versus measured sensible heat flux for the different surfaces, the dual-source model in Norman et al. (1995) had the best agreement with its mean absolute percent difference (MAPD) values being similar to the observational accuracy (i.e., ∼ 20%). From the sensitivity analysis the model appears to have the greatest potential for operational applications since it requires relatively few parameters and is not very sensitive to the uncertainty in most of the model parameters. The single-source models in Kustas et al. (1989) and Troufleau et al. (1996) require accurate estimates of the surface roughness z0m and empirical relationships to account for differences between aerodynamic and radiometric temperature. Therefore, it may be difficult to improve their performance without some independent means of estimating the empirical coefficients and a reliable method for determining z0m. Estimates from the dual-source model in Lhomme et al. (1994) tend to produce the largest scatter with the observations. This may be related to the fact that the model is more sensitive to variations in most of the parameters common in dual-source models. In addition, its response to high surface-air temperature differences in the sensitivity analysis differs from all other models. Therefore, it may be more difficult to obtain reliable estimates from this model on an operational basis.